Feed aggregator

Russell Coker: Finding Corrupt Files that cause a Kernel Error

Planet Debian - Sun, 06/04/2014 - 13:55

There is a BTRFS bug in kernel 3.13 which is triggered by Kmail and causes Kmail index files to become seriously corrupt. Another bug in BTRFS causes a kernel GPF when an application tries to read such a file, that results in a SEGV being sent to the application. After that the kernel ceases to operate correctly for any files on that filesystem and no command other than “reboot -nf” (hard reset without flushing write-back caches) can be relied on to work correctly. The second bug should be fixed in Linux 3.14, I’m not sure about the first one.

In the mean time I have several systems running Kmail on BTRFS which have this problem.

(strace tar cf – . |cat > /dev/null) 2>&1|tail

To discover which file is corrupt I run the above command after a reboot. Below is a sample of the typical output of that command which shows that the file named “.trash.index” is corrupt. After discovering the file name I run “reboot -nf” and then delete the file (the file can be deleted on a clean system but not after a kernel GPF). Of recent times I’ve been doing this about once every 5 days, so on average each Kmail/BTRFS system has been getting disk corruption every two weeks. Fortunately every time the corruption has been on an index file so I don’t need to restore from backups.

newfstatat(4, ".trash.index", {st_mode=S_IFREG|0600, st_size=33, …}, AT_SYMLINK_NOFOLLOW) = 0
openat(4, ".trash.index", O_RDONLY|O_NOCTTY|O_NONBLOCK|O_NOFOLLOW|O_CLOEXEC) = 5
fstat(5, {st_mode=S_IFREG|0600, st_size=33, …}) = 0
read(5,  <unfinished …>
+++ killed by SIGSEGV +++

Related posts:

  1. Bizarre “No space left on device” error from Xen What should have been a routine “remove DIMMs and run...
  2. BTRFS Status March 2014 I’m currently using BTRFS on most systems that I can...
  3. Kernel Security vs Uptime For best system security you want to apply kernel security...
Categories: Elsewhere

Stefano Zacchiroli: historical overview of debian source code

Planet Debian - Sun, 06/04/2014 - 13:19
moar, and moar, and moar debsources stats

A while ago I've announced the availability of ?several stats about Debian source code on http://sources.debian.net. Since then the statistical basis of those stats has increased a lot, and now includes all Debian historical releases, from hamm (July 1998) onward. This allows to appreciate macro-level evolution trends in Free Software, over a period of more than 15 years, through the eyes of a distro that sits at the nice intersection of the eldest, largest, and most reputed distros.

To get there I've added support for sticky suites to the plumbing layer of debsources, and then injected historical releases from http://archive.debian.org. The injection process took about a week (without any sort of parallelism, pretty slow disks, and computing sha256 checksums, ctags, and sloccount on all source files) and has been an "interesting" experience.

When you go back decades in technology time, bit rot is just around the corner, and I've found my share while injecting archive.d.o into sources.d.n. In both cases the respective maintainers (Guillem and Ganneff, kudos) have been positive about and helpful in improving the situation, despite the low impact of the bugs I've found on the average user. That's quite important for the long-term preservation of digital information in general, and for the perennity of access to Free Software in the specific case of Debian.

While we are it, I'm now maintaining a list of bugs affecting sources.d.n but belonging to other packages, in case you fancy helping out but are not a Python hacker. Interestingly enough, quite a bit of those bugs are related to the fact that tools debsources uses (e.g. ctags, sloccount) are also starting to show their age.

You might wander why buzz, rex, and bo are still missing from sources.d.n. That's in fact for similar reasons. Before hamm Debian didn't have complete archive coverage in terms of Sources indexes and .dsc files. Given that debsources rely on both to extract source packages, it first needs to grow an additional abstraction layer that can cope with their absence. It's SMOP, and planned.

And now let's have fun with ctags bombs.

Yours truly,
Stefano “Indiana” Zacchiroli
(credits: KiBi, #debian-ftp)

Categories: Elsewhere

David Herron: Stopping server overload, cleaning up the site front page, disabling comments, and general goodness

Planet Drupal - Sun, 06/04/2014 - 08:09

The last few days the server hosting this site was overloaded, and I finally took a look at the access log, saw a continuous stream of requests that shouldn't be occurring, and realized the "links" row of teasers on the front page needed to go away. The default links row includes one reading "Log in to post comments" but this blog doesn't allow anybody else to register for an account, and in any case comments are handled by Disqus rather than Drupal's commenting system. The link didn't need to be there at all, and the more I looked at the links row the more useless it looked.

Categories: Elsewhere

Keith Packard: Java-Sound-on-Linux

Planet Debian - Sun, 06/04/2014 - 07:30
Java Sound on Linux

I’m often in the position of having my favorite Java program (AltosUI) unable to make any sounds. Here’s a history of the various adventures I’ve had.

Java and PulseAudio ALSA support

When we started playing with Java a few years ago, we discovered that if PulseAudio were enabled, Java wouldn’t make any sound. Presumably, that was because the ALSA emulation layer offered by PulseAudio wasn’t capable of supporting Java.

The fix for that was to make sure pulseaudio would never run. That’s harder than it seems; pulseaudio is like the living dead; rising from the grave every time you kill it. As it’s nearly impossible to install any desktop applications without gaining a bogus dependency on pulseaudio, the solution that works best is to make sure dpkg never manages to actually install the program with dpkg-divert:

# dpkg-divert --rename /usr/bin/pulseaudio

With this in place, Java was a happy camper for a long time.

Java and PulseAudio Native support

More recently, Java has apparently gained some native PulseAudio support in some fashion. Of course, I couldn’t actually get it to work, even after running the PulseAudio daemon but some kind Debian developer decided that sound should be broken by default for all Java applications and selected the PulseAudio back-end in the Java audio configuration file.

Fixing that involved learning about said Java audio configuration file and then applying patch to revert the Debian packaging damage.

$ cat /usr/lib/jvm/java-7-openjdk-amd64/jre/lib/sound.properties ... #javax.sound.sampled.Clip=org.classpath.icedtea.pulseaudio.PulseAudioMixerProvider #javax.sound.sampled.Port=org.classpath.icedtea.pulseaudio.PulseAudioMixerProvider #javax.sound.sampled.SourceDataLine=org.classpath.icedtea.pulseaudio.PulseAudioMixerProvider #javax.sound.sampled.TargetDataLine=org.classpath.icedtea.pulseaudio.PulseAudioMixerProvider javax.sound.sampled.Clip=com.sun.media.sound.DirectAudioDeviceProvider javax.sound.sampled.Port=com.sun.media.sound.PortMixerProvider javax.sound.sampled.SourceDataLine=com.sun.media.sound.DirectAudioDeviceProvider javax.sound.sampled.TargetDataLine=com.sun.media.sound.DirectAudioDeviceProvider

You can see the PulseAudio mistakes at the top of that listing, with the corrected native interface settings at the bottom.

Java and single-open ALSA drivers

It used to be that ALSA drivers could support multiple applications having the device open at the same time. Those with hardware mixing would use that to merge the streams together; those without hardware mixing might do that in the kernel itself. While the latter is probably not a great plan, it did make ALSA a lot more friendly to users.

My new laptop is not friendly, and returns EBUSY when you try to open the PCM device more than once.

After downloading the jdk and alsa library sources, I figured out that Java was trying to open the PCM device multiple times when using the standard Java sound API in the simplest possible way. I thought I was going to have to fix Java, when I figured out that ALSA provides user-space mixing with the ‘dmix’ plugin. I enabled that on my machine and now all was well.

$ cat /etc/asound.conf pcm.!default { type plug slave.pcm "dmixer" } pcm.dmixer { type dmix ipc_key 1024 slave { pcm "hw:1,0" period_time 0 period_size 1024 buffer_size 4096 rate 44100 } bindings { 0 0 1 1 } } ctl.dmixer { type hw card 1 } ctl.!default { type hw card 1 }

As you can see, my sound card is not number 0, it’s number 1, so if your card is a different number, you’ll have to adapt as necessary.

Categories: Elsewhere

Andreas Metzler: exim4 (4.82-6) experimental; urgency=medium

Planet Debian - Sat, 05/04/2014 - 19:00

As GMP has been relicensed to LGPLv3+/GPLv2+ build exim against GnuTLS v3.

Categories: Elsewhere

Mike Hommey: 怒り、失望、ストレス発散

Planet Debian - Sat, 05/04/2014 - 11:21

I started learning japanese calligraphy a few months ago, with no prior experience with a brush and ink. It is an interesting endeavour. For various reasons, I had to skip class for a few weeks, but after the past ten days, I needed some stress relief on paper.


Categories: Elsewhere

Unimity Solutions Drupal Blog: Where do you find the Drupal.org Software working group?

Planet Drupal - Sat, 05/04/2014 - 08:33

Did you know that the Drupal.org Software working group appointed by the Drupal Association provides co

Categories: Elsewhere


Subscribe to jfhovinne aggregator